Page 193 - MEMS Mechanical Sensors
P. 193

182                                                                  Inertial Sensors

                 piezoresistors. A reference structure was used for offset and temperature drift
                 cancellation. The electronic circuitry was operated at a supply voltage of 5V and a
                 supply current of 22 mA. This not only provided filtering, amplification, and buffer-
                 ing of the output signal from the Wheatstone bridge, but also active offset and tem-
                 perature drift compensation by subtracting the output signal of the reference
                 structure and the sensing element. The reported performance of this device was a
                 full-scale measurement range up to ±20G, a resonance frequency of 1.2 kHz, a sen-
                 sitivity of 0.4 mV/V/G with a sensitivity drift of –1.8‰/K and an offset drift of 8
                 µV/V/K.
                    Another example of an integrated piezoresistive accelerometer is described by
                 Reithmueller et al. [16], who developed a similar fabrication technique and experi-
                 mented with different device designs that resulted in comparable performance.



                 8.2.2.2  Capacitive Accelerometers
                 Measuring the displacement of the proof mass capacitively has some inherent
                 advantages over piezoresistive signal pick-off. It provides a large output signal, good
                 steady-state response, and better sensitivity due to low noise performance. The main
                 drawback is that capacitive sensors are susceptible to electromagnetic fields from
                 their surroundings; hence, they have to be shielded carefully. It is also unavoidable
                 that parasitic capacitances at the input to the interface amplifiers will degrade the
                 signal. Usually, a differential change in capacitance is detected. As the proof mass
                 moves away from an electrode, the capacitance decreases, and as it moves towards
                 the electrodes, the capacitance increases. Neglecting fringe field effects, the change
                 in capacitance is given by

                                        1        1                 x
                                                             2
                                                         2
                                                         x <<
                              ∆C =ε A       −          → ε2  A                  (8.5)
                                                            d 0
                                                     
                                     
                                      d −  x  d +                 d  2 0
                                                    x
                                        0
                                                 0
                 which is proportional to the deflection caused by the input acceleration only if the
                 assumption of small deflections is made. For precision accelerometers this assump-
                 tion may be not justifiable, and hence, closed loop control can be used to keep the
                 proof mass deflections small.
                    Early types of capacitive sensors were typically also fabricated by bulk
                 micromachining and the assembling of several wafers by bonding techniques. Most
                 devices had the axis of sensitivity perpendicular to the wafer plane, with the cap
                 wafers on the top and bottom, which, in addition to providing the damping, form
                 the electrodes for capacitive detection. A typical example is shown Figure 8.8 [17].
                    An early, high-precision accelerometer, which can be operated in open and
                 closed loop mode, was reported by Rudolf et al. [18]. The sensor consisted of a sand-
                 wich structure made up from Si-Glass-Si-Glass-Si and is shown schematically in
                 Figure 8.9.
                    The chip size was 8.3 × 5.9 × 1.9 mm with the proof mass size of 4 × 4 × 0.37
                                          –6
                 mm and a mass of 14.7 × 10 kg. The distance of the mass to either electrode at the
                 rest position was 7 µm, which is relatively large; hence, for closed loop opera-
                 tion a voltage of 15V was required. Three <100> silicon wafers were processed, the
                 middle one containing the proof mass and suspension system. These are formed by
   188   189   190   191   192   193   194   195   196   197   198